I am interested in functional connectivity in conditions A and B, and groups 1 and 2. I am looking particularly for group by condition interactions. What I did was 2 functional connectivity analyses per person (ie 2 3ddeconvolves per per person), one for each condition. I did this using a censor with timepoints of interest coded 1, all other TRs coded 0. I used all the output files with 3danova3 to look for main effects and (most importantly) interactions between condition and group.
Here are my main questions: (1) does doing the analysis this way differ significantly from Gang Chen's recommended Context-Dependent Correlation? (http://afni.nimh.nih.gov/sscc/gangc/CD-CorrAna.html) (2) What is the recommended approach to find group by condition interactions in functional connectivity?
Thanks!
~Tali